Hi,
This patch changes a flags argument to spin_lock_irq_save to unsigned long
and removes its unused attribute. The first gets rid of several warnings
and the second is "obviously correct" (at least according to Rusty) :-).
Thanks to DaveM for forcing me to build kernels with a 64 cross compiler :-)
--
Cheers,
Stephen Rothwell [email protected]
http://www.canb.auug.org.au/~sfr/
diff -ruN 2.5.66/include/net/sctp/sctp.h 2.5.66-warnings.1/include/net/sctp/sctp.h
--- 2.5.66/include/net/sctp/sctp.h 2003-03-25 12:08:26.000000000 +1100
+++ 2.5.66-warnings.1/include/net/sctp/sctp.h 2003-03-25 14:19:04.000000000 +1100
@@ -356,7 +356,7 @@
static inline void sctp_skb_list_tail(struct sk_buff_head *list,
struct sk_buff_head *head)
{
- int flags __attribute__ ((unused));
+ unsigned long flags;
sctp_spin_lock_irqsave(&head->lock, flags);
sctp_spin_lock(&list->lock);
From: Stephen Rothwell <[email protected]>
Date: Tue, 25 Mar 2003 14:24:00 +1100
This patch changes a flags argument to spin_lock_irq_save to unsigned long
and removes its unused attribute. The first gets rid of several warnings
and the second is "obviously correct" (at least according to Rusty) :-).
Applied, thank you.
Thanks to DaveM for forcing me to build kernels with a 64 cross compiler :-)
No problem. :)
Stephen Rothwell wrote:
> Hi,
>
> This patch changes a flags argument to spin_lock_irq_save to unsigned long
> and removes its unused attribute. The first gets rid of several warnings
> and the second is "obviously correct" (at least according to Rusty) :-).
>
Hi Stephen. What probably isn't obvious is that the same source is
compiled into a user-level test harness where that variable really _is_
unused. ;-) However, this was just a hack to get rid of a
compilation warning in the testsuite, so I'm glad to see it gone.
thanks!
-jon